Neural Processing Letters

, Volume 11, Issue 1, pp 51–58

Evaluating the Generalization Ability of Support Vector Machines through the Bootstrap

Authors

  • Davide Anguita
    • Dept. of Biophysical and Electronic EngineeringUniversity of Genova
  • Andrea Boni
    • Dept. of Biophysical and Electronic EngineeringUniversity of Genova
  • Sandro Ridella
    • Dept. of Biophysical and Electronic EngineeringUniversity of Genova
Article

DOI: 10.1023/A:1009636300083

Cite this article as:
Anguita, D., Boni, A. & Ridella, S. Neural Processing Letters (2000) 11: 51. doi:10.1023/A:1009636300083

Abstract

The well-known bounds on the generalizationability of learning machines, based on the Vapnik–Chernovenkis (VC) dimension,are very loose when applied to Support Vector Machines (SVMs).In this work we evaluate the validity of the assumption that these bounds are,nevertheless, good indicators of the generalization ability of SVMs.We show that this assumption is, in general, true and assessits correctness, in a statistical sense, on several pattern recognition benchmarks throughthe use of the bootstrap technique.

bootstrapgeneralizationsupport vector machinesVC dimension
Download to read the full article text

Copyright information

© Kluwer Academic Publishers 2000